Cadmium induces Wnt signaling to upregulate proliferation and survival genes in sub-confluent kidney proximal tubule cells

Cycloheximide and actinomycin D were purchased from Roth (Karlsruhe, Germany), lactacystin, rabbit serum (cat. #S2632) and protease inhibitor cocktail were obtained from Sigma-Aldrich (Deisenhofen, Germany). Recombinant protein G sepharose 4B was from Invitrogen (Karlsruhe, Germany). Antibodies (all mouse monoclonal) used were: anti-β-catenin (cat. # 610153; 1:1000) and anti-E-cadherin (cat. #610181; 1:2500) from BD Biosciences (Heidelberg, Germany), anti-α-catenin (cat. # ab19446; 1:1000) from Abcam (Cambridge, UK), anti-CHOP (clone9C8) (cat. # MA1-250; 1:1000) from Dianova GmbH (Hamburg, Germany), anti-TCF4 (clone6H5-3) (cat. # 05-511; 1:1000) from Millipore GmbH (Schwalbach, Germany), anti-GAPDH (cat. # G8795; 1:20,000), anti-β-actin (cat. # A5316; 1:10,000) and anti-γ-tubulin (cat. # T6557; 1:10,000) from Sigma-Aldrich. The secondary sheep anti-mouse horseradish peroxidase (HRP)-linked IgG (NA931; 1:5000) was purchased from Amersham Biosciences (Freiburg, Germany). All other chemicals were of the highest purity grade available.

Cells from the S1-segment of rat PT (WKPT-0293 Cl.2) were immortalized by retroviral transfection with SV40 large T-antigen [18] and cultured as previously described [19] at 37°C in 5% CO₂. Apart from some ECIS experiments (see below), all Cd²⁺ experiments were performed in serum-free medium (SFM). Unless otherwise indicated, 25 μM Cd²⁺ was used. For immunoblot experiments comparing subconfluent and confluent cells ~50% confluence was obtained by seeding 5 × 10⁵ cells per well and 100% confluence by seeding 1 × 10⁶ cells per well in 6-well plates. After 48 hours WKPT-0293 Cl.2 cells were incubated in the absence or presence of Cd²⁺ in SFM before harvesting.

The protocol was adapted from the method described by Andrews and Faller [20]. Cells were harvested into 400 μl of ice cold buffer A (10 mM HEPES-KOH, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM DTT, 0.05% Nonidet P-40, 0.2 mM Pefabloc SC and a protease inhibitor cocktail), and allowed to lyse on ice for 10 min. After a pre-run at 4°C at 500 × g for 1 min to remove unlysed cells and debris, further centrifugation was performed at 16,000 × g for 1 min. The supernatant containing cytosolic proteins was collected. Exactly 100 μl of buffer B (20 mM HEPES-KOH, pH 7.9, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM DTT, 0.05% Nonidet P-40, 0.2 mM Pefabloc SC and protease inhibitor cocktail), was added to the pellet and incubated on ice for 20 min. The suspension was centrifuged at 16,000 × g for 2 min and the supernatant containing the nuclear proteins was collected.

Protein concentration of samples was determined by the Bradford method [21], using bovine serum albumin as a standard. SDS-PAGE and transfer were performed exactly as described elsewhere [17]. Signals were quantified using ImageJ or TINA 2.09 software. For comparisons between control and experimental conditions specific signals were first normalized to loading markers (β-actin, γ-tubulin or GAPDH).

Cells (5 × 10⁴) were grown in 24-well plates on glass cover-slips. After treatment coverslips were washed with PBS, fixed with 4% paraformaldehyde in PBS, permeabilized with 0.2% TritonX-100 for 5 min, and blocked with 1% BSA and 100 μg/ml RNAse A for 2 h. Thereafter, the cells were incubated with primary antibodies against β-catenin (1:1000), TCF4 (1:250) or E-cadherin (1:1000) followed by staining with anti-mouse Alexa-Fluor 488 conjugated secondary antibody (1:400) (Invitrogen GmbH; Karlsruhe, Germany) for 1 h and subsequent propidium iodide (PI) staining (10 μg/ml) for 5 min. Then the cells were washed and samples were visualized with a LEICA TCS SP5 confocal laser microscope (Wetzlar, Germany) by exciting at 488 nm (emission 520 nm) to detect β-catenin and TCF4, and at 543 nm (emission 617 nm) for PI. E-cadherin overexpressing and mock-transfected cells were viewed with a Zeiss Axiovert 200 M microscope (Carl Zeiss, Jena, Germany) using filters for FITC with excitation/emission wavelengths of 480/535 nm, respectively. Images were acquired at fixed exposure times (600 ms), processed, and analyzed semiquantitatively with MetaMorph software (Universal Imaging Corporation, Downingtown, PA).

Nuclear fractions were used for EMSA of TCF4 binding to cyclin D1 and ABCB1 promoters. Based on the consensus TCF/LEF-binding motifs, CTTTGA/TA/T [22] the wild-type TCF/LEF binding sequence of the human cyclin D1 promoter 5'-CTCTGCCGGGCTTTGATCTTTGCTTAACA-3' [23] and the binding sequence of human ABCB1 5'-GGCTTTGAA GTATGA-3' [24] were selected. They were end-labeled with [γ-³²P] dATP by incubating oligodeoxyribonucleotide strands with 5× reaction buffer and 10 U T4 polynucleotide kinase (Fermentas, St. Leon-Rot, Germany) for 1 h at 37°C. Then labeled oligonucleotides were allowed to anneal at room temperature for 10 min and 20 μg protein from each sample was used in 25 μl binding reactions, which consisted of 1 μg poly dI-dC, in 5× binding buffer (50 mM Tris HCl;pH 8.0, 750 mM KCl, 2.5 mM EDTA, 0.5% Triton-X 100, 62.5% glycerol (v/v) and 1 mM DTT). To determine specificity of DNA binding, samples were incubated with or without 20 ng of unlabeled competitor DNA for 10 min at room temperature. Then 0.1 ng of labeled probe was added and samples were further incubated for 20 min at room temperature. Samples were separated on a 5% non-denaturing polyacrylamide gel in 0.5% TBE and visualized by autoradiography.

Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany) and first strand cDNA was synthesized with the Omniscript RT kit (Qiagen), using 1 μg of RNA per 20 μl reaction and oligo(dT) primer. cDNA was then utilized in PCR reactions for c-Myc and GAPDH as previously described [17]. Remaining primers are summarized in the Additional file 1.

Following activation at 95°C for 15 min, the following PCR conditions for each primer pair were performed: For Lef-1, TCF-3, TCF-4, E-cadherin and β-catenin 28 cycles at 94°C for 20 s, 62°C for 30 s, 72°C for 45 s; for cyclin D1 34 cycles at 94°C for 30 s, 56°C for 60 s, 72°C for 90 s; for Abcb1a 32 cycles at 95°C for 80 s, 56°C for 60 s and 72°C 90 s; for CHOP 28 cycles at 94°C for 30 s, 62°C for 30 s and 72°C for 45 s.

Myc-epitope-tagged full length human TCF4 and a deletion mutant lacking the NH₂-terminal 30 amino acids (ΔN-hTCF4), which are essential for its interaction with β-catenin, were inserted in the expression vector pcDNA3 [25] and were a gift of Dr. Bert Vogelstein (Baltimore, Maryland, USA). Full-length E-cadherin inserted into the plasmid pL31NU [26] was a gift of Dr. Rolf Kemler (Freiburg Germany), the full length mouse CHOP construct in pcDNA1 was a gift of Prof. David Ron (New York, NY, USA) and the ΔN-β-catenin construct in pCGN lacking the NH₂-terminal stretch of 132 amino acids required for its degradation [27] was a gift of Prof. Avri Ben-Ze'ev (Rehovot, Israel).

Unless otherwise indicated, WKPT-0293 Cl.2 cells (5 × 10⁵ cells per well in 6-well plates) were transiently transfected 24 h post-seeding with the various constructs using Lipofectamine 2000 reagent (Invitrogen) following manufacturer's instructions. Plasmids and Lipofectamine reagent were used at a constant ratio of 1:2.5 and incubated for up to 36 h.

Wnt signaling was assessed using the well-described TOPflash assay [25]. Briefly, 5 × 10⁴ cells per well grown for 24 h in 24-well plates were transiently transfected with TOPflash or FOPflash (Millipore) using poly(ethyleneimine) (PEI; Sigma) from a stock solution of 1 mg/ml at a ratio of 1:0.75 (volume DNA:volume PEI). For experiments with ΔN-β-catenin, cells were first transfected with ΔN-β-catenin construct and then with TOPflash or FOPflash 6 h later. Luciferase activity was determined using Tropix Luminescence assay kit (Applied Biosystems Applera Deutschland GmbH, Darmstadt, Germany) after 24 h and the measurements were performed in a Mithras LB 940 multimode microplate reader (Berthold Technologies GmbH, Bad Wildbad, Germany). The readings were normalized against protein concentration determined by the Bradford method [21] for each sample. For experiments with ~50% confluent or 100% confluent cells 5 × 10⁴ and 9 × 10⁴ cells per well, respectively, were grown in 24-well plates. The degree of cell confluence refers to the time of transfection with TOPflash or FOPflash.

The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method was modified as previously described [17] and measured on a Helios Epsilon spectrophotometer (Thermo Scientific, Langenselbold, Germany). The values were normalized to the control, which was equivalent to 100% cell viability, to determine cell death rates. To account for the loss of cells during the transfection procedure, different cell numbers (1.5 × 10⁴ for empty vector and CHOP, 2.5 × 10⁴ for E-cadherin in 48 well plates) were plated to obtain a similar cell density prior to treatment with Cd²⁺.

Eighteen to 24 hours after transfection, WKPT-0293 Cl.2 cells were trypsinized, centrifuged at 400 × g for 5 min, and plated in ECIS 8W10E cell culture arrays at 3.5 - 5.5 × 10⁵ cells per well. For experiments evaluating the effect of Cd²⁺ on the attachment and proliferation of CHOP-transfected cells serum-containing medium (SCM) included 50 μM CdCl₂. The changes in resistance (R) and capacitance (C) of the recording electrodes were determined using an ECIS™1600R instrument (Applied BioPhysics, Troy, NY), as previously described [17]. C_{40 kHz} and R_{400 Hz} values reflect electrical properties of epithelial monolayers due to cell attachment/spreading/proliferation and barrier formation, respectively. Whereas C_{40 kHz} mirrors attachment and spreading of cells R_{400 Hz} values are indicative of the epithelial barrier integrity, especially establishment of tight intercellular contacts. For experiments evaluating cell attachment and proliferation, the cells were incubated in SCM, with twice daily medium renewal. Once the cells had established a functional monolayer with stable R and C readings, they were washed once with SFM and the medium was replaced with SFM ± 20 μM Cd²⁺.

The nonfluorescent, dye 5-(and-6)-carboxy-2',7'-dichlorodihydrofluorescein diacetate (carboxy-H₂DCFDA) is oxidized to fluorescent carboxydichlorofluorescein by H₂O₂, peroxyl radicals and peroxynitrite anion (Invitrogen GmbH; Karlsruhe, Germany). WKPT-0293 Cl.2 cells (2 × 10⁵; 6-well plates) were loaded with 20 μM carboxy-H₂DCFDA, in the presence of the ABCB1 inhibitor PSC833 (1 μM) to improve dye loading, for 30 min at 37°C. After washing, cells were gently trypsinized, pelleted by centrifugation at 400 × g for 3 min, brought into suspension, and added to each well of a 96 well plate ± Cd²⁺. Data were collected every 60 sec at λ_ex/λ_em of 500/535 nm on a Berthold Mithras LB940 fluorescence microplate reader. When used, cells were pre-incubated with 100 μM α-tocopherol for 1 h prior to dye loading and remained in the presence of α-tocopherol for the duration of the measurements.

One early event associated with Cd²⁺ nephrotoxicity is disruption of the homophilic E-cadherin interaction [15, 16]. In the present study, exposure of sub-confluent WKPT-0293 Cl.2 cells to Cd²⁺ (25 μM in SFM) for 3-9 h led to cytoplasmic and nuclear accumulation of β-catenin (Fig. 1A), indicating disruption of the AJ complex. Statistical analysis of 3 different experiments displays significant β-catenin increase in cytosol already after 3 h and in nuclei after 6 h. This was corroborated by confocal immunofluorescence microscopy of Cd²⁺ treated PTC which exhibited a more diffuse β-catenin distribution in the cytoplasm and nucleus compared to the controls where β- catenin was predominately found at the cell borders (Fig. 1B).

A possible explanation for the higher levels of β-catenin is increased synthesis induced by Cd²⁺. However, mRNA levels of β-catenin were unchanged by Cd²⁺ (Additional file 2a) and the translation inhibitor cycloheximide had no effect on Cd²⁺-induced β-catenin redistribution (Additional file 2b). To find out whether the increase in β-catenin might be due to inhibition of the proteasomal degradation complex by Cd²⁺, we utilized lactacystin, a potent proteasomal inhibitor. After 6 h the increase of the cytosolic and nuclear pool of β-catenin observed with Cd²⁺ was not mimicked by lactacystin (Additional file 2c), indicating that Cd²⁺ does not inhibit the proteasome. Thus the elevation of β-catenin must be a consequence of redistribution of β-catenin from the membrane pool to cytosol and nuclei.

Alpha-catenin is another member of the AJ complex assembly, which mediates interaction between AJs and actin filaments, therefore disruption of the complex should also lead to the release of α-catenin. Similarly to β-catenin, α-catenin content in the cytoplasm and nuclei steadily increased under Cd²⁺ treatment of sub-confluent WKPT-0293 Cl.2 cells at 3-9 h (Fig. 1C). β-catenin/TCF/LEF1 mediated transcription has been reported to be inactivated through β-catenin binding to E-cadherin [28] or α-catenin [29]. However, binding of β-catenin to E-cadherin as well as α-catenin is reduced by Cd²⁺ exposure (Fig. 1D), supporting the disruptive action of Cd²⁺ on the AJ and negating the probability of transcriptional inactivation of β-catenin by α-catenin.

The pivotal transcription factor for the canonical Wnt pathway, the TCF4 protein, is responsible for transactivation of cell proliferation and survival genes. Strikingly, we observed increased localization of TCF4 in nuclei upon Cd²⁺ exposure from an early time point of 3 h and remained elevated up to 9 h (Fig. 2A). This could also be visualized in confocal laser scanning micrographs (data not shown). A prerequisite for activation of the Wnt pathway-mediated expression of target genes is binding of β-catenin to TCF4 in the nucleus. To evidence this, immunoprecipitation was performed with β-catenin and bound TCF4 was determined by immunoblotting. Cd²⁺-treated samples exhibited more β-catenin-TCF4 binding as early as 3-6 h (Fig. 2B). To confirm β-catenin/TCF4 interaction, we repeated the experiment at 3-9 h Cd²⁺ incubation by pulling down TCF4 and immunoblotting for β-catenin (Additional file 3a) thus establishing that there is significantly more β-catenin bound to TCF4 under Cd²⁺ treatment conditions.

To assess the activation of the Wnt transcriptional machinery upon Cd²⁺ exposure, we performed luciferase activity measurements using the TOPflash reporter construct, which contains optimal TCF binding sites, in sub-confluent WKPT-0293 Cl.2 cells. Cd²⁺ maximally increased TOPflash activity starting at 3-6 h but terminated at 12 h. After 6 h, the TCF reporter activity was 1.81 ± 0.21-fold (means ± SEM of 8 experiments; P = 0.009) (Fig. 2C). FOPflash activity, a negative control, was not affected by Cd²⁺ (data not shown). For comparison, transfection of PTC with a ΔN β-catenin, which is resistant to proteasomal degradation [30] resulted in a 5.7 ± 1.5 fold (means ± SEM of 4 experiments; P = 0.02) increase of luciferase activity, compared to the control vector (Fig. 3B). The expression of Wnt signaling target genes involved in cell proliferation and survival, c-Myc, cyclin D1 and Abcb1a, was also increased by Cd²⁺ exposure for up to 12 h (Fig. 2D). Experiments using EMSAs showed binding of TCF4 to the promoter sequence of both ABCB1 (Fig. 2E) and cyclin D1 (Additional file 3b) that was enhanced by Cd²⁺ for up to 9 h, but peaked at 3 h. In order to assure that β-catenin was responsible for the transcription of target genes via the β-catenin/TCF4 transcription factor complex, we over-expressed wild-type hTCF4 and mutant ΔN-hTCF4. This mutant form of TCF4 lacks the NH₂-terminal region required for interaction with β-catenin [25]. Cd²⁺ (25 μM for 3-6 h) increased the expression of c-Myc in wild-type TCF4 over-expressing cells more profoundly than in vector-transfected or ΔN-hTCF4 mutant transfected cells (Fig. 2F). This confirms the role of β-catenin in mediating the up-regulation of Wnt target genes under Cd²⁺ exposure of kidney PTC.

Interestingly, we found that the Wnt response activated by Cd²⁺ is dependent on the confluence and proliferation status of the cells. With 100% confluent cells, we did not detect β- and α-catenin translocation to the cytosol and nuclei of Cd²⁺-exposed PTC, in contrast to ~50% confluent cells (Fig. 3A). This was associated with markedly reduced TOPflash activity. As shown in Fig. 3B, there was increased nuclear β-catenin bound to TCF4 in ~50% confluent cells but not in 100% confluent cells exposed to Cd²⁺. Moreover, ~50% confluent cells exhibited an ~2-fold increase in TOPflash transcriptional activity upon exposure to 25 μM Cd²⁺ for 6 h (P = 0.007), whereas 100% confluent cells were not responsive (Fig. 3C). We also observed higher E-cadherin expression in 100% confluent than in ~50% confluent cells (Fig. 3D). Since β- and α-catenin are associated with E-cadherin at the AJ, it could account for the observed dependency of Cd²⁺ to induce Wnt signaling on confluence.

To test the hypothesis that high levels of E-cadherin could prevent nuclear β-catenin translocation and Wnt signaling induced by Cd²⁺, we overexpressed full-length E-cadherin in WKPT-0293 Cl.2 cells. E-cadherin expression was increased by ~70% and was found to be uniformly distributed at the cell borders (Additional file 4a). When cells were subjected to 25 μM Cd²⁺ for 6 h, the increase in cytoplasmic and nuclear β-catenin induced by Cd²⁺ was abolished in E-cadherin overexpressing cells (Fig. 4A).

Because Cd²⁺-induced β-catenin translocation to the nucleus triggers β-catenin/TCF4-mediated Wnt signaling and increases expression of cell proliferation genes (see Fig. 2), we investigated the impact of E-cadherin on cell proliferation by ECIS. The ECIS technique measures the electrical properties of epithelial monolayers due to cell attachment/spreading/proliferation and barrier formation. C_{40 kHz} reflects attachment, spreading and proliferation of cells whereas R_{400 Hz} replicates epithelial barrier integrity and establishment of tight intercellular contacts (see Methods). C_{40 kHz} for control cells decreased from values of the cell-free electrode to its baseline values (8-10 nF) after ~20 h (Fig. 4B), reflecting the establishment of an intact monolayer with coverage of the whole electrode surface. In contrast, E-cadherin over-expression in WKPT-0293 Cl.2 cells significantly retarded cell proliferation. Interestingly, however, once a stable monolayer was established, E-cadherin over-expressing cells were less sensitive to Cd²⁺ (Fig. 4C). The decrease of R_{400 Hz}, indicative of a disruption of the epithelial barrier, became apparent within hours after addition of 20 μM Cd²⁺ in serum-free medium in vector-transfected cells, but proceeded far more slowly in E-cadherin over-expressing cells. Corresponding ΔC_{40 kHz} increased by 1.76 ± 0.20 nF/h in vector-transfected cells as opposed to 0.91 ± 0.35 nF/h in E-cadherin-transfected PTC (means ± SEM of 3 experiments) reflecting increased detachment of cells from the electrode. This suggested that E-cadherin over-expression not only protects the epithelial monolayer from Cd²⁺-induced disruption of cellular junctions but also from detachment and death.

This was further tested by investigating the impact of E-cadherin overexpression on cell viability of WKPT-0293 Cl.2 cells exposed to Cd²⁺. Transfected cells were treated with 25 μM Cd²⁺, assessed using the MTT assay. Following treatment with Cd²⁺, cell viability decreased by ~30% in vector-transfected cells. Conversely, Cd²⁺ could decrease cell viability by only ~9% in E-cadherin over-expressing cells (P < 0.01) (Fig. 4D). Taken together, these data indicate that E-cadherin overexpression decreases Wnt signaling-mediated cell proliferation and in confluent PTC monolayers is protective against Cd²⁺-induced disruption of the cellular junctions, detachment and cell death.

ROS can be generated by Cd²⁺ [31] and are known to disrupt TJs and AJs resulting in decreased transepithelial electrical resistance and redistribution of β-catenin [32]. Thus, we investigated whether the β-catenin accumulation in cytosol and nuclei is ROS-dependent. As shown in Fig. 5A, Cd²⁺ (25 μM) caused an increase in ROS within minutes after Cd²⁺ exposure which was determined by oxidation of the fluorescent probe carboxy-H₂-DCFDA. ROS formation steadily increased over time reaching 30 ± 9% above controls at 2 h exposure (P < 0.01; n = 3), which was prevented by the antioxidant α-tocopherol (100 μM) (Fig. 5A). But Cd²⁺ exposure showed similar effects on β-catenin distribution into the cytosol of PTC whether in the presence or absence of α-tocopherol (Fig. 5B), indicating that β-catenin release from the membrane and E-cadherin-bound pool to the intracellular space is ROS-independent.

CHOP, also known as GADD153, is a pro-apoptotic protein that is involved in the endoplasmic reticulum (ER) stress pathway. Increased expression of CHOP has been previously reported in the presence of Cd²⁺ in kidney PTC as a function of ROS formation [33]. Moreover, CHOP has been previously shown to negatively regulate the Wnt signaling pathway through binding to TCF4 [34]. We also found that Cd²⁺ treatment (25 μM) of WKPT-0293 Cl.2 cells increased the expression of CHOP in a time dependent manner (up to 18 h) (Fig. 5C), which was abrogated in the presence of α-tocopherol (Fig. 5D). CHOP was pulled down with TCF4 in immunoprecipitation experiments confirming CHOP-TCF4 interaction. This was further enhanced by Cd²⁺ exposure (25 μM for 6 h) (Fig. 5E). Next, we investigated the expression of the Wnt target gene c-Myc in cells over-expressing CHOP. When treated with Cd²⁺, CHOP-overexpressing cells still showed an increment of c-Myc expression, indicating that CHOP does not prevent Cd²⁺-induced Wnt signaling (Fig. 5F). Cell viability studies with Cd²⁺ for 6 h (Additional file 4b) or 12 h (data not shown) showed no effect of CHOP overexpression on Cd²⁺-induced cell death in PTC. Similarly, measurements of cell proliferation demonstrated that CHOP overexpression had no effect on proliferation in Cd²⁺-exposed cells (data not shown). Both findings are in accordance with the c-Myc expression data (Fig. 5F). From these data, we concluded that CHOP plays no significant role in preventing cell proliferation and survival of PT cells following Cd²⁺ exposure though CHOP does bind to TCF4 (see Fig. 5E).

A possible reason for the discrepancy between our data and the literature may be the striking observation of increased total TCF4 elicited by Cd²⁺ exposure from an early time point of 3 h up to 9 h Cd²⁺ at the protein (Fig. 6A) and mRNA (Fig. 6B) levels. Moreover, TCF4 up-regulation by Cd²⁺ in sub-confluent cells was strongly reduced in confluent monolayers (Fig. 6C). To strengthen our observation that TCF4 up-regulation by Cd²⁺ is at the transcriptional level, we used the transcriptional inhibitor actinomycin D. In the presence of actinomycin D (10 μg/ml, pre-incubated for 1 h), Cd²⁺ treatment failed to up-regulate TCF4 protein, when compared to cells treated with Cd²⁺ alone (Fig. 6D). Hence increased TCF4 expression may override CHOP inhibition.

In contrast to the effect of Cd²⁺ on TCF4, additional transcriptional components of Wnt/β-catenin signaling were not up-regulated by Cd²⁺ exposure in sub-confluent PTC. TCF3 mRNA levels were similar in control and Cd²⁺ treated cells for up to 18 h (Additional file 4c; TCF3). LEF1, a marker of mesenchymal cells [35], could not be detected, indicating the epithelial characteristic of the cell line and absence of any epithelial-to-mesenchymal transformation (EMT) phenomena upon Cd²⁺ treatment (data not shown). Similarly, no change was detected in E-cadherin mRNA expression which is suppressed during EMT via the Wnt pathway (reviewed in [36]) (Additional file 4c; E-cadherin).